Mx using different Panel Calculation Model

[HoshangMustafa]

Hi,

I tried two models with different Panel Calculation Model: one with Partial stiffening, and the other Without stiffening. I got two different diagrams for Mx, the former with uniform diagram and the latter with uniform varying diagram. Please find the attached pictures. I wonder why?

[Krzysztof_Wasik]

HI @HoshangMustafa 

You have used

in model 1

shell model for slab which gives expected "quasi linear" moment distribution. Moment changes "lineary" because it is transferred by several nodes along beam length

and

in model 2

Rigid diaphragm for slab which do not transfer torsion moment to side beams . In second model moment is transferred to side beam only by  beam 18 (middle beam so between endnos and middlenode torsion moment value is constant.)

Results are as expected.

[HoshangMustafa]

The model with partial stiffening seems like hand calculation. So my intent is to make the results being similar (using fe and partial stiffening). How one can make them identical?

[Krzysztof_Wasik]

Hi @HoshangMustafa 

Remark

You have used fully rigid slab calculation model in your model (not partial stiffening).

I am aftraid I do not understand your question. Please explain your expectations with more details.

Different slab calculation models gives different load disribution and different results. Results for slab FE model will be not the same as for slab simulated by fully rigid diaphragm.

Giving user possibility to choose method (calculation model) corresponding to his expectations is the main purpose for their various calculation model in the software. You should choose the model which corresponds to your needs.

You cannot expect the same results for those two (FE and Fully rigid) slab models. Torsion moment in the beam is the effect of bending in slab simulated by FE, while bending is neglected in fully rigid model.

[HoshangMustafa]

@Krzysztof_Wasik 

You have used fully rigid slab calculation model in your model (not partial stiffening).

I'm sorry. It's my mistake. I've incorrectly labelled it Rigid diaphragm. If you check the settings in Panel Calculation Model of my model, you will notice that.

[Krzysztof_Wasik]

Hi @HoshangMustafa 

I am sorry I have not noticed that.

Rigid diaphragm does not transfer torsionat all (I should have noticed that)

It does not change anything regarding Mx. When Finite Elements are not defined in the slab, torsion is transferred to beams  11 12 25 26  (in their middle points) only by beams  18 28

Slabs transfers only vertical forces to supporting beams.

[HoshangMustafa]

Hi @Krzysztof_Wasik 

You wrote:

When Finite Elements are not defined in the slab, torsion is transferred to beams  11 12 25 26  (in their middle points) only by beams  18 28

But My for Beam 18 is -11.48 while Mx (torsion) for beam 12 is 5.74 for Case DL2 (11.48/2=5.74). Why is it torsion for 12 half the bending moment for 18? I think 5.74 is reaction of the girder for applied moment of 11.48?

I tried modelling it with Cladding. It seems results (Mx and My for beam 12) using Cladding and Partial stiffening are identical for gravity loads. Am I right?

[Krzysztof_Wasik]

Hi @HoshangMustafa 

I understand it in the following way (please correct me if I have missed your point)

• Bending moment in beam endnode 18 - My= 11.30 kN*m
• It is transferred to beam 12 as torsion Mx in node 18
• It is reflected as difference in Mx torsion between beam 12 left part , Mx=5.65, and its right part Mx=-5.65.
• 11.30 = 5.65-(-5.65) so in my opinion results are correct. 

Claddings and partial stiffening calculation model generates the same load distribution.

• cladding is weightless while for slab with partial stiffening model, slab selfweight can be considered
• cladding is not rigid in horizontal plane (while slab with partial stiffening is rigid in horizontal plane)

[HoshangMustafa]

Hi @Krzysztof_Wasik 

You wrote:

for slab with partial stiffening model, slab selfweight can be considered

The Mx results (values) for beam 12 is nearly identical for FE & partial stiffening for cases DL2 & LL1. The big difference is for case DL1 (self-weight). I wonder how classical method of trapezoidal and triangular been used for analysis, especially for self-weight case. Am I correct?

[Krzysztof_Wasik]

Hi @HoshangMustafa 

I have compared maximum mx value in beam 12 for caes DL1 Dl2 LL1 in both model.

For all cases difference is about 30 %. So differences are similar for all cases

When only distributed load on slab would be considered (neglected loads applied to beam elements) difference is 32 % regardless of applied load value

Selfweight is distributed for partial stiffening in the following way.

The load in the middle is duplicated.

distributed loads values

[HoshangMustafa]

Hi @Krzysztof_Wasik 

You wrote:

Selfweight is distributed for partial stiffening in the following way.

The load in the middle is duplicated.

DL2 & LL1 are also distributed this way. The load in the middle is duplicated.

You wrote:

bending is neglected in fully rigid model (post 4)

Rigid diaphragm does not transfer torsion at all (post 6)

In Full restrained model, for case DL2 I noticed that shear (Fz) is small for beam 12 compared to Cladding and Partial restrained model. I wonder why is that. Is it related to your posts 4 & 6?

[HoshangMustafa]

Hi @Krzysztof_Wasik 

I tried checking this simple structure playing around with Reduced Results for Panel 17 MRy Cut 2-4 Case 2 and Mx for beam 9.

MRy: 9.49kNm

Mx: 3.97kNm

Is it comparable?

[Krzysztof_Wasik]

Hi @HoshangMustafa 

I cannot obtain values described in your email (for panel 17 beam 9 and case 2) using model shared previously.

Example Mx moment for beam 9 (2.66-2.48 instead of 3.97)

Could you please share model with those values displayed (and screenshots presenting what you are comparing)

[HoshangMustafa]

Hi @Krzysztof_Wasik 

Thanks. I forgot to state that I modified the model for simplicity. Please find the attached link.

I don't know where one can download Autodesk castscreen app?

[Krzysztof_Wasik]

Hi @HoshangMustafa 

I think that you should compare  value from panel cut (9.49) with difference between moment Mx from both ends of beam 9 (3.97-(3.97))=7.94. Missing part (about 1.5) is transferred by other bars attached to nodes 2 and 4

Results matches better when slab is separated from beams 11 12 (no load transfer from slab to other bars attached to nodes 2 and 4), then 

Mry =14.57 in slab

7.25- (-7.27) = 14.52 - Mx difference in 9 between nodes 2 and 4

[HoshangMustafa]

Hi @Krzysztof_Wasik 

You wrote:

I think that you should compare  value from panel cut (9.49) with difference between moment Mx from both ends of beam 9 (3.97-(3.97))=7.94. Missing part (about 1.5) is transferred by other bars attached to nodes 2 and 4

How one can display this 1.5 moment transferred by other bars attached to nodes 2 and 4?

You wrote:

Results matches better when slab is separated from beams 11 12 (no load transfer from slab to other bars attached to nodes 2 and 4)

How one can separate slab from beams 11 12?

[Krzysztof_Wasik]

Hi @HoshangMustafa 

Ad1 (Missing 1,5)

You have to check balance of nodal moments in all bars connected to beam 9 ends.

Ad 2 (separating panel from beams)

I have defined pabel width slightly smaller than beam 9 length.

[HoshangMustafa]

Hi @Krzysztof_Wasik 

You wrote:

You have to check balance of nodal moments in all bars connected to beam 9 ends.

I tried but couldn't manage this. Can you elaborate more?

You wrote:

I have defined pabel width slightly smaller than beam 9 length.

Yes, it works.

[Krzysztof_Wasik]

Hi @HoshangMustafa 

Please share verification you have made. I will try to verify it

[HoshangMustafa]

Hi @Krzysztof_Wasik